Hi Developers, I have committed a few new feature in the Release_2_1-maint branch of mailman CVS. Please export/checkout the new source code and test them in your test/working mailman installation and feed back. Also, I want to ask 'nimda.txt' in the tests/msgs directory is of any use in the mailman source code. None of the test scripts refer to this file. I even did "find . | xargs grep nimda" only to get "Binary file ./tests/msgs matches". If none of you can find usage of this file, I want to remove from the source tree because some companies prohibit to download a file which contain nimda file. (I know it's nonsense but ...) Here is the diff for current NEWS file. (I believe Barry can fix my poor English.) More will come later because I get requests of merging patches personally. But, please be patient as I will check the patches and bugs for assigning to myself when it comes your turn. ;-) Cheers, -- Tokio Kikuchi, tkikuchi@ is.kochi-u.ac.jp Update of /cvsroot/mailman/mailman In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv8029 Modified Files: Tag: Release_2_1-maint NEWS Log Message: update NEWS for new features and bug fixes Index: NEWS =================================================================== RCS file: /cvsroot/mailman/mailman/NEWS,v retrieving revision 2.43.2.28 retrieving revision 2.43.2.29 diff -u -d -r2.43.2.28 -r2.43.2.29 --- NEWS 18 May 2004 02:15:55 -0000 2.43.2.28 +++ NEWS 16 Oct 2004 04:54:33 -0000 2.43.2.29 @@ -6,7 +6,32 @@ 2.1.6 (XX-XXX-200X) - - Bugs and patches: 955381 (older Python compatibility). + - List owners can now set how many days to hold the messages in moderator + request queue. cron/checkdb automatically discard old messages. (790494) + + - Improved mail address sanity check. (1030228) + + - SpamDetect.py now checks attatchment header. (1026977) + + - Subject_prefix can be configured to include sequential number which + is taken from post_id variable. Also, the prefix is always put at + the start of the subject, i.e. like "[list-name] Re: original subject" + + - Now, list owners can use Scrubber to get the attachments scrubbed + (held in the web archive), if the site admin permits it in mm_cfg.py. + Also, introduced are SCRUBBER_DONT_USE_ATTACHMENT_FILENAME and + SCRUBBER_USE_ATTACHMENT_FILENAME_EXTENSION in Defaults.py for scrubber + behavior. (904850) + + - Filter attachments by filename extensions. (1027882) + + - Bugs and patches: 955381 (older Python compatibility), 1020102/1013079/ + 1020013 (fix spam filter removed), 665569 (newer postfix bounce + detection), 970383 (moderator -1 admin requests pending), 873035 + (subject handling in -request mail), 799166/946554 (makefile + compatibility), 872068 (add header/footer via unicode), 1032434 + (KNOWN_SPAMMERS check for multi-header), 1025372 (empty Cc:), + 789015 (fix pipermail URL), 2.1.5 (15-May-2004)
At 2:52 PM +0900 2004-10-16, Tokio Kikuchi wrote:
Here is the diff for current NEWS file. (I believe Barry can fix my poor English.) More will come later because I get requests of merging patches personally. But, please be patient as I will check the patches and bugs for assigning to myself when it comes your turn. ;-)
While we're at it, can we update the documentation so that we can deprecate FAQ entries 5.8 and 5.12? It sure would be nice to not have so many people asking about support questions with Mailman 2.1.5 using older versions of Python, or MTAs that don't implement VERP.
Also, if we're going to require VERP support on the MTA with 2.1.5 and beyond, can we go ahead and turn on "VERP_CONFIRMATIONS = Yes" in mm_cfg.py, so that we can also eliminate all the questions regarding the funky subject lines which led to FAQ 4.52?
On a separate matter, the way that full personalization works is really, really annoying. When enabling this on lists I manage, I've gotten more complaints than on all other items combined on those same lists.
I'm going to work out a way to hack out the most offensive things and submit a patch, but has anyone else had issues with it? Am I totally out in left field, or does someone else already have a patch that I haven't seen?
-- Brad Knowles, brad@stop.mail-abuse.org
"Those who would give up essential Liberty, to purchase a little temporary Safety, deserve neither Liberty nor Safety."
-- Benjamin Franklin (1706-1790), reply of the Pennsylvania
Assembly to the Governor, November 11, 1755SAGE member since 1995. See http://www.sage.org/ for more info.
On 10/15/2004 22:52, "Tokio Kikuchi" tkikuchi@is.kochi-u.ac.jp wrote:
Also, I want to ask 'nimda.txt' in the tests/msgs directory is of any use in the mailman source code. None of the test scripts refer to this file. I even did "find . | xargs grep nimda" only to get "Binary file ./tests/msgs matches". If none of you can find usage of this file, I want to remove from the source tree because some companies prohibit to download a file which contain nimda file. (I know it's nonsense but ...)
My guess is that the nimda.txt file is left over and can be removed without harm.
You might want to refer folks who want to run test "virus" messages through their Mailman system ("system" = the MTA and its filtering and Mailman and whatever else is involved at a given site) to http://www.eicar.org (note...eicar.com is very different).
By including the reference, you avoid having anything in the distribution which triggers (sensible) virus scanners.
--John (who just used eicar yesterday in another context)
On Sat, 16 Oct 2004 08:52:15 -0700 John W. Baxter wrote:
...[snip]...
You might want to refer folks who want to run test "virus" messages through their Mailman system ("system" = the MTA and its filtering and Mailman and whatever else is involved at a given site) to http://www.eicar.org (note...eicar.com is very different).
They are??? www.eicar.org and www.eicar.com look identical (and have the same IP address).
By including the reference, you avoid having anything in the distribution which triggers (sensible) virus scanners.
--John (who just used eicar yesterday in another context)
Hi,
Brad Knowles wrote:
At 2:52 PM +0900 2004-10-16, Tokio Kikuchi wrote:
Here is the diff for current NEWS file. (I believe Barry can fix my poor English.) More will come later because I get requests of merging patches personally. But, please be patient as I will check the patches and bugs for assigning to myself when it comes your turn. ;-)
While we're at it, can we update the documentation so that we candeprecate FAQ entries 5.8 and 5.12? It sure would be nice to not have so many people asking about support questions with Mailman 2.1.5 using older versions of Python, or MTAs that don't implement VERP.
Can you write a patch for INSTALL/README and upload it to SF patch tracker? It would be highly appreciated. I myself updated one of my server's sendmail to use 2.1.5.
Also, if we're going to require VERP support on the MTA with 2.1.5and beyond, can we go ahead and turn on "VERP_CONFIRMATIONS = Yes" in mm_cfg.py, so that we can also eliminate all the questions regarding the funky subject lines which led to FAQ 4.52?
Sounds reasonable. Anyone against this?
On a separate matter, the way that full personalization works isreally, really annoying. When enabling this on lists I manage, I've gotten more complaints than on all other items combined on those same lists.
I'm going to work out a way to hack out the most offensive thingsand submit a patch, but has anyone else had issues with it? Am I totally out in left field, or does someone else already have a patch that I haven't seen?
-- Tokio Kikuchi, tkikuchi@ is.kochi-u.ac.jp http://weather.is.kochi-u.ac.jp/
Also, if we're going to require VERP support on the MTA with 2.1.5and beyond, can we go ahead and turn on "VERP_CONFIRMATIONS = Yes" in mm_cfg.py, so that we can also eliminate all the questions regarding the funky subject lines which led to FAQ 4.52?
Sounds reasonable. Anyone against this?
Oops. This is only for "invite". Looks like patch is there, so I should commit into the CVS.
-- Tokio Kikuchi, tkikuchi@ is.kochi-u.ac.jp http://weather.is.kochi-u.ac.jp/
At 10:29 AM +0900 2004-10-17, Tokio Kikuchi wrote:
While we're at it, can we update the documentation so that we candeprecate FAQ entries 5.8 and 5.12? It sure would be nice to not have so many people asking about support questions with Mailman 2.1.5 using older versions of Python, or MTAs that don't implement VERP.
Can you write a patch for INSTALL/README and upload it to SF patch tracker? It would be highly appreciated.
I'll try to do this within the next few days, if someone else doesn't beat me to it.
-- Brad Knowles, brad@stop.mail-abuse.org
"Those who would give up essential Liberty, to purchase a little temporary Safety, deserve neither Liberty nor Safety."
-- Benjamin Franklin (1706-1790), reply of the Pennsylvania
Assembly to the Governor, November 11, 1755SAGE member since 1995. See http://www.sage.org/ for more info.
Hi,
Anyone cares about 1029275 (Maximum number of members per list) ? This patch was submitted at sf and here a month ago and still nothing. I just want to read "we don't want this feature" or "it's a bad patch, it's needs more work here and there". Kikuchi ?
Thank you for keep contributors motivated.
Best regards, Gustavo Franco
On Oct 16, 2004, at 1:52 AM, Tokio Kikuchi wrote:
Here is the diff for current NEWS file. (I believe Barry can fix my poor English.) More will come later because I get requests of merging patches personally. But, please be patient as I will check the patches and bugs for assigning to myself when it comes your turn. ;-)
Well, I'm not Barry, but I'll checked your English and it looks good to me, in case you want another opinion.
- Bugs and patches: 955381 (older Python compatibility), 1020102/1013079/
1020013 (fix spam filter removed), 665569 (newer postfix bouncedetection), 970383 (moderator -1 admin requests pending), 873035(subject handling in -request mail), 799166/946554 (makefilecompatibility), 872068 (add header/footer via unicode), 1032434(KNOWN_SPAMMERS check for multi-header), 1025372 (empty Cc:),789015 (fix pipermail URL),
If Barry pushes through the documentation updates I made, we'll also have fixed 948152 (Out of date link on Docs webpage) but I don't know if this should go in the NEWS file or no since it's not exactly a bug related to the release. We should remember to mark this bug as fixed, though, whenever the update happens.
Terri
On Sat, 2004-10-16 at 01:52, Tokio Kikuchi wrote:
I have committed a few new feature in the Release_2_1-maint branch of mailman CVS. Please export/checkout the new source code and test them in your test/working mailman installation and feed back.
Awesome, and thanks!
Also, I want to ask 'nimda.txt' in the tests/msgs directory is of any use in the mailman source code. None of the test scripts refer to this file. I even did "find . | xargs grep nimda" only to get "Binary file ./tests/msgs matches". If none of you can find usage of this file, I want to remove from the source tree because some companies prohibit to download a file which contain nimda file. (I know it's nonsense but ...)
Yes, you can get rid of it. It was there once to test the email parser, but it's no longer necessary.
Here is the diff for current NEWS file. (I believe Barry can fix my poor English.) More will come later because I get requests of merging patches personally. But, please be patient as I will check the patches and bugs for assigning to myself when it comes your turn. ;-)
Thanks -- I've made some minor corrections to the file. I appreciate all your help in working on 2.1.6.
-Barry
On Sat, 2004-10-16 at 02:22, Brad Knowles wrote:
While we're at it, can we update the documentation so that we can deprecate FAQ entries 5.8 and 5.12? It sure would be nice to not have so many people asking about support questions with Mailman 2.1.5 using older versions of Python, or MTAs that don't implement VERP.
As to 5.8 and 4.42, I do consider the fact that Mailman 2.1.5 breaks under older Python's a bug that should be fixed in 2.1.6 (and was fixed in CVS very early on). Mailman 2.1.x should still run under Python 2.1. I do however /recommend/ using Python 2.3.
As for 5.12, I think it's an oversight on my part that the bounce probe feature requires VERP support in the MTA. I would like to see the probe become optional in 2.1.6 so that if you do not have VERP, Mailman will process bounces in the 2.1.4 way (i.e. not send a probe). I don't know if that's feasible or even desirable -- how much pain is it for mailman-users that VERP is required? From 5.12, it sounds like those using Mailman in some hosting arrangements are prevented from upgrading because of this. That's unfortunate.
Also, if we're going to require VERP support on the MTA with 2.1.5 and beyond, can we go ahead and turn on "VERP_CONFIRMATIONS = Yes" in mm_cfg.py, so that we can also eliminate all the questions regarding the funky subject lines which led to FAQ 4.52?
/Should/ we require VERP? My preference would be "no".
On a separate matter, the way that full personalization works is really, really annoying. When enabling this on lists I manage, I've gotten more complaints than on all other items combined on those same lists.
Can you provide more detail about those complaints? What don't people like about it? The only difference between "personalized" messages and "fully personalized" messages is the To header, so if they don't like that, you shouldn't fully personalize the messages.
-Barry
On Sat, 2004-10-16 at 02:22, Brad Knowles wrote:
While we're at it, can we update the documentation so that we can deprecate FAQ entries 5.8 and 5.12? It sure would be nice to not have so many people asking about support questions with Mailman 2.1.5 using older versions of Python, or MTAs that don't implement VERP.
Oh, BTW, could you (or someone) update those FAQ entries to explain my intent about Python backward compatibility? Thanks,
-Barry
On Sat, 2004-10-16 at 23:48, Gustavo Franco wrote:
Anyone cares about 1029275 (Maximum number of members per list) ? This patch was submitted at sf and here a month ago and still nothing. I just want to read "we don't want this feature" or "it's a bad patch, it's needs more work here and there". Kikuchi ?
I'm not personally in favor of adding this feature to the 2.1 line. I understand that because 2.2 will never come out and I'm way behind on 3.0 that pressure builds to add new features to the 2.1 point releases. I'd still like to resist that pressure as much as possible, but I can support getting some new features (in a backward compatible way) into 2.1.(x>5). Let's take them on a case-by-case basis (and Tokio has a lot of sway here because he's doing the hard work ;).
-Barry
At 9:07 PM -0400 2004-10-17, Barry Warsaw wrote:
As to 5.8 and 4.42, I do consider the fact that Mailman 2.1.5 breaks under older Python's a bug that should be fixed in 2.1.6 (and was fixed in CVS very early on). Mailman 2.1.x should still run under Python 2.1. I do however /recommend/ using Python 2.3.
Fair enough. We can update the FAQ Wizard entry when 2.1.6 ships.As for 5.12, I think it's an oversight on my part that the bounce probe feature requires VERP support in the MTA. I would like to see the probe become optional in 2.1.6 so that if you do not have VERP, Mailman will process bounces in the 2.1.4 way (i.e. not send a probe). I don't know if that's feasible or even desirable -- how much pain is it for mailman-users that VERP is required? From 5.12, it sounds like those using Mailman in some hosting arrangements are prevented from upgrading because of this. That's unfortunate.
I don't know how much work it would require to make this turned on by default, but to allow fallback behaviour if not supported. The alternative is to revert to the older behaviour and turn off the VERP features by default. I'm not fond of that alternative, but it may be better for the broader Mailman community, especially those at hosting sites where more modern MTA features may not be available.
/Should/ we require VERP? My preference would be "no".
If that's the direction you prefer, that's fine. My point is that we should align all our features, and if we're going to require VERP in one area, we might as well make use of it in another, especially if doing so will help to reduce a very common complaint.
Can you provide more detail about those complaints? What don't people like about it? The only difference between "personalized" messages and "fully personalized" messages is the To header, so if they don't like that, you shouldn't fully personalize the messages.
The munging of the To: and Cc: headers, throwing away all previous information that was there, etc... has been the major complaint that has been voiced to me. My understanding is that some of the template/personalization features are not enabled unless you turn on "full personalization", and if that's the case, we need a fourth alternative.
If I'm wrong, then we might want to be more clear in our documentation, so that people do not needlessly turn on full personalization when they don't have to do so in order to get the features they're looking for.
-- Brad Knowles, brad@stop.mail-abuse.org
"Those who would give up essential Liberty, to purchase a little temporary Safety, deserve neither Liberty nor Safety."
-- Benjamin Franklin (1706-1790), reply of the Pennsylvania
Assembly to the Governor, November 11, 1755SAGE member since 1995. See http://www.sage.org/ for more info.
At 9:08 PM -0400 2004-10-17, Barry Warsaw wrote:
Oh, BTW, could you (or someone) update those FAQ entries to explain my intent about Python backward compatibility? Thanks,
If this hasn't already been done, I will do so.-- Brad Knowles, brad@stop.mail-abuse.org
"Those who would give up essential Liberty, to purchase a little temporary Safety, deserve neither Liberty nor Safety."
-- Benjamin Franklin (1706-1790), reply of the Pennsylvania
Assembly to the Governor, November 11, 1755SAGE member since 1995. See http://www.sage.org/ for more info.
At 9:08 PM -0400 2004-10-17, Barry Warsaw wrote:
Oh, BTW, could you (or someone) update those FAQ entries to explain my intent about Python backward compatibility? Thanks,
Okay, 5.8 and 5.12 have been updated with notes at the end, based on comments made in this thread. Please let me know if you think there need to be changes made, or if you think they look okay.
-- Brad Knowles, brad@stop.mail-abuse.org
"Those who would give up essential Liberty, to purchase a little temporary Safety, deserve neither Liberty nor Safety."
-- Benjamin Franklin (1706-1790), reply of the Pennsylvania
Assembly to the Governor, November 11, 1755SAGE member since 1995. See http://www.sage.org/ for more info.
Barry Warsaw wrote:
As for 5.12, I think it's an oversight on my part that the bounce probe feature requires VERP support in the MTA. I would like to see the probe become optional in 2.1.6 so that if you do not have VERP, Mailman will process bounces in the 2.1.4 way (i.e. not send a probe).
Before I start working on this, does anyone have a patch? I think we should introduce BOUNCER_USE_VERP_PROBE (or something) and set its default to No.
Also, if we're going to require VERP support on the MTA with 2.1.5 and beyond, can we go ahead and turn on "VERP_CONFIRMATIONS = Yes" in mm_cfg.py, so that we can also eliminate all the questions regarding the funky subject lines which led to FAQ 4.52?
/Should/ we require VERP? My preference would be "no".
OK Barry, we should step back. I want however to introduce patch 923428 so site admin can use full feature of VERP_CONFIRMATIONS.
-- Tokio Kikuchi, tkikuchi@ is.kochi-u.ac.jp http://weather.is.kochi-u.ac.jp/
On Sun, 17 Oct 2004 21:12:54 -0400, Barry Warsaw barry@python.org wrote:
On Sat, 2004-10-16 at 23:48, Gustavo Franco wrote:
Anyone cares about 1029275 (Maximum number of members per list) ? This patch was submitted at sf and here a month ago and still nothing. I just want to read "we don't want this feature" or "it's a bad patch, it's needs more work here and there". Kikuchi ?
I'm not personally in favor of adding this feature to the 2.1 line. I understand that because 2.2 will never come out and I'm way behind on 3.0 that pressure builds to add new features to the 2.1 point releases. I'd still like to resist that pressure as much as possible, but I can support getting some new features (in a backward compatible way) into 2.1.(x>5). Let's take them on a case-by-case basis (and Tokio has a lot of sway here because he's doing the hard work ;).
I understand your point and Tokio replied me at sf. Do you think that is a feature that can be included on 3.0 ? I can keep extending it over 2.1 line (updating 1029275) and add it to 3.0 tree. I'm not into mailman project to deep (yet) but i'll read that page about mm3 at zope.org now.
What's your opinion on distributors adding new features to mailman 2.1 packages? Considering that 3.0 won't be released soon is a good idea announce that hot new features won't be added if not really needed on mm2.1 tree, isn't it? jmo.
Thanks, Gustavo Franco -- stratus@acm.org
On 10/16/2004 14:13, "David Relson" relson@osagesoftware.com wrote:
On Sat, 16 Oct 2004 08:52:15 -0700 John W. Baxter wrote:
...[snip]...
You might want to refer folks who want to run test "virus" messages through their Mailman system ("system" = the MTA and its filtering and Mailman and whatever else is involved at a given site) to http://www.eicar.org (note...eicar.com is very different).
They are??? www.eicar.org and www.eicar.com look identical (and have the same IP address).
Oops...they are the same.
This was Safari helpfully completing a URL it had seen before within eicar.org, and not completing it--not having see it--in eicar.com.
The proper reference would probably be the more complete
eicar.org/anti_virus_test_file.htm or eicar.com/anti_virus_test_file.htm
Thanks for objecting! ;-)
--John
participants (7)
-
Barry Warsaw -
Brad Knowles -
David Relson -
Gustavo Franco -
John W. Baxter -
Terri Oda -
Tokio Kikuchi